文献类型：专利

英文题名：Single edge based dynamic sampled unbalance data classifying system, has non-balance system for determining utilization rate of network structure, updating unit updating claim value, and classification module classifying unknown sample

作者：FAN Q;GAO D;LI D;WANG Z

机构：[1]UNIV EAST CHINA SCI & TECHNOLOGY

申请号：CN105787046-A

申请日：2016-02-28

公开日：2016-07-20

语种：英文

收录：DERWENT

摘要：NOVELTY - The system has a non-balance system for determining utilization rate of a network structure according to balance data set scale. The non-balance system optimizes a network mode with learning rate, charge factor and maximum iteration times using gradient descending process. An updating unit updates each claim value. A resulting classification module classifies unknown samples. A network machine is provided with a network layer and a nerve cell data layer. The nerve cell data layer is mutually connected with a node. USE - Single edge based dynamic sampled unbalance data classifying system. ADVANTAGE - The system performs down-sampling process and classifier training process to achieve dynamic sampling data collection function so as to avoid loss of information. The system effectively processes classification unbalanced data. DESCRIPTION OF DRAWING(S) - The drawing shows a flowchart illustrating operation of a single edge based dynamic sampled unbalance data classifying system. '(Drawing includes non-English language text)'